GIS, Big Data and Lessons from John Snow
Doug Specht*1
1
University of Westminster, Communication and Media Research Institute.
January 07, 2015
Summary
This paper examines the work of Snow within the ultra-modern context of big data and GIS, and
questions the results that may be born from GIS and Big Data alone. Arguing that while GIS and
spatial research have a great potential for unearthing trends, caution must be taken to ensure we do
not generate dangerously misleading information about geographical and sociological connections.
The paper concludes that when reflecting on contemporary GIS and Big Data practice, that we should
look to the work behind Snow’s map, rather than being besotted by his famous geographic
visualisation.
KEYWORDS: Big Data, GIS, John Snow, data visualisation, data analysis
The Big Data age is unquestionably here. In recent years, the volume of data collected and stored by
business and government organizations has snowballed. Driven by reduced costs of storage and ever
increasing analysis capabilities Big Data has become a big industry. The benefits of Big Data have also
been widely reported, the McKinsey Global Institute (MGI) has cited examples of the transformative
effect of Big Data from sectors as dispersed as health care to retail to manufacturing to political
campaigns (Manyika et al., 2011). Research conducted at the Massachusetts Institute of Technology
shows that companies that use “data-directed decision making” enjoy a 5% – 6% increase in
productivity, and that mass analysis of mobile phone calling patterns can help detect flu outbreaks
(Boyd, & Crawford, 2012; Kirkpatrick, 2013). Big Data has also been put to use in humanitarian
efforts, researchers from Sweden’s Karolinska Institute and Columbia University have used data from
Digicel, Haiti’s largest cell phone provider, to determine the movement of displaced populations after
the 2010 earthquake, aiding in the distribution of resources; Tweets in Indonesia have been analysed to
predict how people fare with food price volatility, allowing pre-emptive measures to be put in place to
reduce scarcity shock. And social media output anlysis in the United States and Ireland have shown to
be good early indicators of spikes in unemployment (Kirkpatrick, 2013.
It is the combination of this rich data source, collected through mobile devices, coupled with the Big
Data Paradigm - which promises to turn ever larger and imperfect, complex, often unstructured data
into actionable information, within a discourse of increased speed, efficiency and inclusivity (Hilbert,
2013; Burns, 2014) - that presents exciting opportunities for geographers, GIS analysts and the GI
industry as a whole. By 2020, more than 70 percent of mobile phones are expected to have GPS
capability, up from 20 percent in 2010 (Hilbert, 2013), leading to a massive increase to the flood of
spatially located data already generated every day. The Big Data Paradigm suggests that the more data
we have, the better our predictive modelling and analytics will be. Caution, however, must be taken in
our rush to exploit these new vast pools of data and information for predictive analysis.
The evidence for a more nuanced approach to combining Big Data and GIS comes from a surprising
point in mapping history, namely the unpicking of the mythology surrounding John Snow and his
infamous 19th Century Cholera maps. Through an exploration of John Snow’s methods and the
*
[email protected]
application of the principles of critical, participatory and feminist GIS, this paper seeks to ground our
understanding of the value of Big Data in GIS and remind us of the under lying principles of GIS
analysis.
Snow’s study of the ‘‘Broad Street outbreak’’ has long been heralded as the start of spatioanylitical
research and is oft cited as a fundamental example of epidemiology and medical geography (Koch &
Denike, 2009). Our fascination with the map however, has somewhat distorted our understanding of
the methods employed in its creation, and the conclusions that where drawn at the time. Two
misconceptions persist around Snow’s maps that have implications for how we merge GIS and Big
Data. Firstly, that it was his maps that led Snow to reach the conclusion that Cholera was water born.
And Secondly, that his maps provided good evidence for this conclusion.
When Snow first presented his hypothesis to the parish officials that the water pump maybe the source
of the Broad Street outbreak he neither presented a map, nor did he allude to the idea that a map had
been instrumental in his discovery (Bordy et al., 2000). Moreover, the first edition of On the mode of
communication of cholera, published in 1849, contained no maps, it was not until 1854 that his spot
map was first published, possibly due to the influence of Shapter, whom Snow had cited in OMCC’s
second edition. It would appear then that Snow had developed and tested his hypothesis well before he
drew his map. This is not an unlikely scenario given that he was already engaged in an ambitious study
of cholera in South London. It was likely these earlier studies that led him to conclude that a sharp
localised outbreak pointed to a contaminated pump rather than, as commonly reported, an induction
arrived at primarily from the geographical facts of the case (Brody et al., 2000). Snow’s map then did
not give rise to the insight, but was the tool used to confirm and illustrate an already held hypothesis
and conclusion.
More widely known about Snow’s work is that, despite his body of evidence, he had a hard time
convincing those around him that the water pump was the source of contamination, leading to the
emotive myth of him striding into Broad Street and breaking off the water pumps handle. Snow’s
contemporaries, his readers and the parish had been unconvinced by his arguments. It was not so much
his theory to which they objected, but rather his lack of detailed consideration of other potential sources
of contagion (Koch & Denike, 2009). It was this lack of support for his ideas that led Snow to draw his
maps, to prove and illustrate his theory. Simply plotting deaths on a map, however, did not lead others
to reach the same conclusions, nor the immediate, unquestioning adoption of his theory (Brody et al.,
2000). “On examining the map given by Dr Snow, it would clearly appear that the centre of the outburst
was a spot in Broad-street, close to which is the accused pump; and that cases were scattered all round
this nearly in a circle, becoming less numerous as the exterior of the circle is approached. This certainly
looks more like the effect of an atmospheric cause than any other” was the conclusion reached by
Edmund A Parkes in his review of On the mode of communication of cholera (Parkes, 1855: 458 cited
in Brody et al., 2000). Indeed spot maps such as Snow’s had previously been used by both contagionists
and anticontagionists to advance their stance in Yellow Fever research as early as 1798 (Brody et al.
2000). Snow’s map alone was not enough to convince either his contemporaries or other parties as to
his, albeit correct, theory.
What can we learn from Snow and his approach in the age of GIS and Big Data, and how to we integrate
this in our work? The trap of examining Snow’s work is in positioning him as a mapmaker without an
appreciation of Snow the thinker. Data are meaningful because of how someone collects, interprets,
and forms arguments with it. Data are not neutral. The goal of Snow’s maps was not to produce data,
but to filter data, to reframe his hypothesis. Snow’s original theorem was discounted, his map was
misinterpreted, but together they were able to support each other, leading to the saving of many lives.
The Big Data age is moving us into a new epistemology in which knowledge politics are deeply
embedded in what counts as data, how those data are represented and the proposes for which those
representations may be used. Evidence has shown that we will continue to see marginal increases in
predictive performance even to a massive scale (Junqué de Fortuny, 2013), but these increases are
useless if we are representing this data through a epistemology that embodies an unquestioned form of
legitimacy and power, such as that often given to map-based artefacts, and which leads to wide
misinterpretation or which is based on unsound hypothesis. Big Data, plotted in the form of a highly
sophisticated and accurate map, may easily seduce us into concluding that we have learned something.
Big Data should be considered as less a physical entity and more as a paradigm shift. It won’t solve
issues on its own, but it allows us to think differently. Big Data thinking opens our view to nontraditional data for predictive analytics, but we must be careful as we embrace this technological idea
not to become besotted by it. Key to the integration of Big Data as a meaningful tool of analysis is
ensuring that Big Data is viewed as much as a social innovation as a technical one. The future of GIS
isn’t about pulling mass amounts of data towards the centre like a black hole, it’s about pushing for
simple, purposeful and considered data collection leading to better collaboration, analysis and
understanding. While our understanding of Snow’s map production and methodology might be skewed
by myth and legend, the truth of the story is still a shining example of how to blend mapped data,
visualisation and theory. The Broad Street maps and Snow’s work demonstrate that the mere act
arranging data graphically in space does not yield new understanding without the support of a solid,
considered and researched theory, and that our theories will be widely dismissed if we have not explored
all hypotheses and represented them appropriately, regardless of the size of our Big Data set.
Figure 1: Original map made by John Snow in 1854. Cholera cases are highlighted in black. Published by C.F. Cheffins, Lith,
Southhampton Buildings, London, England, 1854 in Snow, John. On the Mode of Communication of Cholera, 2nd Ed, John
Churchill, New Burlington Street, London, England, 1855. This image is in the public domain due to its age
References
Boyd, D., & Crawford, K. (2012). Critical questions for big data: Provocations for a cultural,
technological, and scholarly phenomenon. Information, Communication & Society, 15(5), 662-679.
Brody, H., Rip, M. R., Vinten-Johansen, P., Paneth, N., & Rachman, S. (2000). Map-making and mythmaking in Broad Street: the London cholera epidemic, 1854. The Lancet, 356(9223), 64-68.
Burns, R. (2014). Rethinking big data in digital humanitarianism: practices, epistemologies, and social
relations. GeoJournal, 1-14.
Hilbert, M. (2013). Big data for development: From information-to knowledge societies. Available at
SSRN 2205145.
Junqué de Fortuny, E., Martens, D., & Provost, F. (2013). Predictive Modelling With Big Data: Is
Bigger Really Better?. Big Data, 1(4), 215-226.
Kirkpatrick, R. (2013). Big Data for Development. Big Data, 1(1), 3-4
Koch, T., & Denike, K. (2009). Crediting his critics' concerns: Remaking John Snow's map of Broad
Street cholera, 1854. Social science & medicine, 69(8), 1246-1251.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., & Byers, A. H. (2011). Big
data: The next frontier for innovation, competition, and productivity.
Biography
Doug Specht is a Doctoral Researcher at the University of Westminster where he is exploring how
digital media and GIS are used in legitimizing and codifying local knowledge within the context of
International Development. He is also the Director of VOZ, a PGIS platform that supports human and
environmental rights.